From what I can tell, it's a combination of a few useful tools:
1) Configuration monitoring – if an app is running on six identical servers, and all of the sudden one box has a different config file sitting on it, or somebody installs updates that they weren't supposed to, I'd like an email about that.
2) A "source of truth" for your cloud resources. Smarter people than I have articulated our need for this at my employer, but the simplest way I know of to describe it is that you can query the configuration of your boxes to figure things out. "Show me all the boxes running Nginx 1.6," for example, or "Show me all the boxes that haven't gotten the latest Bash update." Not sure how it does that, but it seems to be in there.
I don't think AWS Config handles either of your examples. It records the state of various AWS resources, not the state of the software/environment running within an EC2 instance.
Measuring your entire server architecture involves many things from software versions provided by the package maintainer (Ubuntu/Debian, RHEL/CentOS, custom, etc), memory capacity of your machines, disk allocation, and so forth.
Changes to this server architecture -- anything and everything -- should be managed in a controlled environment.
AWS Config provides insight into this, notifying you of when any changes occur.
They provide you with a constant stream of changes occurring, like getting Tweets whenever anything changes.
Amazon already has AWS CloudWatch, this takes it a step further than monitoring performance, resource utilization, and operational health. It delves into what happens before performance thresholds are breached and operational health is diminished.
The most interesting thing about it is that it provides an event bus that can stream inventory updates (via SQS).
This is super useful for cloud management platforms that keep a copy of you resources in Amazon e.g. for policy enforcement and compliance. Now the copy of the data will be much closer to the actual state of the resources in the cloud.
The ManageIQ project already added support for it via this pull request:
It would be interesting to look through the AWS service health reports for unrelated services to determine if it suffers more problems as a result of this.
10 comments
[ 1.3 ms ] story [ 33.1 ms ] thread1) Configuration monitoring – if an app is running on six identical servers, and all of the sudden one box has a different config file sitting on it, or somebody installs updates that they weren't supposed to, I'd like an email about that.
2) A "source of truth" for your cloud resources. Smarter people than I have articulated our need for this at my employer, but the simplest way I know of to describe it is that you can query the configuration of your boxes to figure things out. "Show me all the boxes running Nginx 1.6," for example, or "Show me all the boxes that haven't gotten the latest Bash update." Not sure how it does that, but it seems to be in there.
Changes to this server architecture -- anything and everything -- should be managed in a controlled environment.
AWS Config provides insight into this, notifying you of when any changes occur.
They provide you with a constant stream of changes occurring, like getting Tweets whenever anything changes.
Amazon already has AWS CloudWatch, this takes it a step further than monitoring performance, resource utilization, and operational health. It delves into what happens before performance thresholds are breached and operational health is diminished.
It's an auditing tool, basically.
This is super useful for cloud management platforms that keep a copy of you resources in Amazon e.g. for policy enforcement and compliance. Now the copy of the data will be much closer to the actual state of the resources in the cloud.
The ManageIQ project already added support for it via this pull request:
https://github.com/ManageIQ/manageiq/pull/1049
Also see the following blog post by yours truly
http://www.redhat.com/en/about/blog/red-hat-brings-expanded-...
It would be interesting to look through the AWS service health reports for unrelated services to determine if it suffers more problems as a result of this.
http://i.imgur.com/fUT6ROI.png